Page Operations
Browse SpaceQ. Why is a global network of collections and services registries needed?
We think that global registries are needed for three reasons: (1) to improve the discoverablity of value-added information resources that are described according to formal collection description principles; (2) enhance the functions of scholarly communications infrastructures (repositories, data centers, library catalog's etc.) to better support international research collaborations; and (3) provide the technical foundations for a new generation of registry applications that seamlessly interoperate with mainstream Web 2.0 applications.
Q. Who would benefit from a global network of registries?
We envisage that a global network of collections and services registries will initially benefit three communities: (1) collecting organizations (such as libraries, museums, archives and higher education and research institutions) seeking to improve public access search services, at the same time improving their internal collection management infrastructures (2) research communities seeking information about datasets, data collections, publications and other information relevant to their research interests; and (3) researchers, research managers, research funders, governments, and the general public interested in research trends and outcomes.
Q. How would a global network of registries help researchers?
A global network of interoperable collections and services registries would help researchers by contributing to the new cyberinfrastructure fabric currently in development, such as the Object Reuse and Exchange (ORE) initiative. It would do this by implementing international standards for describing digital collections and developing a data-model/schema for sharing collection-level metadata between current collections/services registry projects. Taken together, the capacity to create consistent collection-level metadata, and then sharing it between interoperable registries, would enable the development of 'meta-registries' that would provide online access much larger pool of collections records, resources and services. Furthermore, machine-to-machine Web services offered by these meta-registries would enable a wide range of downstream value-added services to users.
It would also help by providing researchers and research groups with consistent frameworks, standards and tools to help them organize their value-added information resources into 'collections' (whether they call them collections, or not, is irrelevant) and to make these available to others worldwide. Some domain-specific research collections are currently being made available to registered users via bespoke search engines; however, although useful, this information is generally not aggregated at higher levels, nor does it appear in the results of search engines like Google. Indeed, the proposed solution opens the possibility for Google-like services for research collections. The goal here would be to facilitate interdisciplinary research and stimulate innovation through the serendipitous re-use of research materials in new contexts. Arguably, domain-specific pools of resources of international scope would provide a much stronger motivation to scholars to use and contribute.
Q. Are there good community models to draw on?
Yes. The Open Language Archives Community (OLAC) have adopted an international standards to describe research collections related to audio recordings and publications related to linguistics, musicologists and ethnographers. These collections and repositories can be located online using the OLAC search engine.
Q. Why isn't Google search good enough for the task?
While it is generally acknowledged Google search provides an excellent discovery service to locate individual digital 'items', it does not provide filtered, collection-level, search results for users (nor arguably could it using its current page-rank technologies). However, nothing in principle precludes Google from indexing collection-level metadata and complementing it with their item-level search results to provide new search services. Indeed, online library search services like OCLC WorldCat and Libraries Australia are already 'mashing up' collection- and item-level information from their catalogs with links to Google search, Amazon.com and so on.
Q. What standards would be adopted?
Two standards are under consideration: ISO2146 for a registry data model/schema and Dublin Core Collections Application Profile for collection-level metadata. The registry services would adopt the Web Services standards maintained by W3C and OASIS. There is no intention to standardize registry software itself; rather, a 'reference' registry schema/data model based on ISO2146 could be adopted, along with registry service interfaces and API's for software application developers. Other standards and protocols may need to be developed to enable this proposal to be implemented.
Q. Why a global initiative?
When scaled internationally, an information infrastructure that allows researchers, and the general public, to discover research collections is a simple but powerful tool. This is particularly the case with research datasets which are still largely invisible to most generic WWW search systems. However, there is no internationally implemented standardized description for these materials, and real world standardization initiatives generally occur within particular disciplinary domains, with their own local indexes or registries. However, given that research practice is increasingly an international, multidisciplinary, enterprise there is limited motivation to register with or search in local registries. Hence, we are proposing an international alliance of locally-maintained registries but which are linked to national and international meta-registries. How such an alliance would be governed, or funded, are open questions for discussion at the meeting.
Q. What is a 'collection'?
We acknowledge that discussions about the definition of a 'collection' can quickly degenerate into 'collection-metaphysics-talk'; that is, essentially irresolvable arguments over the normative definition, scope and granularity of digital collections. In contrast, we use the definition of a collection adopted by the Dublin Core Collections Application Profile; that the term collection "can be applied to any aggregation of physical and/or digital resources". In practice, guidelines would be needed to address issues related to the granularity of collections and the structural relationships between super- and sub-collections.
Q. What is a Service?
Dublin Core Collections Application Profile defines a service as a "system that provides access to the Items within the Collection". In this context, data centers, digital libraries, and institutional repositories aggregate scholarly collections within systems that support sophisticated access methods, for example, OPeNDAP, SRU/SRW, and OAI-PMH. Such "services" enable access to, searching over, and harvesting from collections within a research repository. A collections/service registry aims to record in standardized machine-readable form the descriptions of these services.